Merge Strategies and World Views

One of the books I have been reading most recently is called “Traffic” by Tom Vanderbilt. The subtitle is “Why We Drive The Way We Do (and what it says about us)”.

This is not the kind of book you would normally think of as an entertaining read. But like most people who drive a decent amount I have some fairly strong opinions about driving habits, rules and customs. So clearly this is a topic that promises to be of interest to someone like me.

“Traffic” takes a very interesting (and detailed) look at some of these habits, rules and customs, and analyzes some of the (often mistaken) assumptions we make about things like speeding, congestion, lane changing, traffic safety, and so on.

A word of warning: if you are looking for support for a pet theory (such as “large trucks cause lots of accidents” or “slow drivers are more dangerous than fast ones on 4 lane highways”) you probably won’t find it here. Generally speaking this book is an examination of the statistical evidence that supports such theories, and for the most part the simplistic theories that you and I have – especially if you have contrarian views – are not supported by the evidence.

One interesting suggestion the author makes right near the beginning of the book has to do with the practice of “late merging”.  This is what you do when you come to a “merge left” sign (usually because of construction or an accident on the highway). 

 There are three possibilities at such a sign: early merge (get over as quickly as possible), gradual merge (merge gradually, but don’t wait until the end of the lane), and late merge (go right up to the end of the disappearing lane and merge at the last minute). 

The most socially acceptable method of merging in such a situation is the early merge. We Canadians, being the polite people we are, seem to consider it almost our duty to get over as soon as possible – what you might call “extreme early merging”.

This is not so much a matter of traffic efficiency or even driving etiquette as it is a reflection of our character. You know – that guy who busts past you in the right lane and expects other people to let him in ahead of the rest of us obedient drivers. Isn’t that queue jumping?

Well, it turns out that according to Vanderbilt, studies have been done (of course they have!) that show it is more efficient to late merge. If you go right up to the end of the lane and then merge into the adjacent lane you win on two counts. First, you maximize the available space by keeping the disappearing lane full. And second, there is much less jockeying and anticipating as you look for someone to let you in.

In other words, late merging is more efficient. And it would be even more efficient (and safer) if drivers just accepted the same kind of rule they accept at a four-way stop. Stop, let somebody from the other lane go, then go.

That’s pretty simple, but it’s not the rule we use because, as Vanderbilt says, “there seems to be a whole worldview contained in each of the merge strategies that have been tried.” And many people simply cannot get over the attitude that “I’m acting like a selfish SOB if I push right up to the end of the line.”

I found this especially interesting because of another contrarian merge strategy I use every time I merge onto the highway just outside of our town – often several times a week. It’s what I call the “inside move”, or in this context might be called a “delayed merge”.

The scenario is this: a long two-lane on-ramp merges from the right with the regular three lanes of the highway. Well back on the ramp people are told a merge is coming and they obediently line up in the left lane. Not only does this slow the entire left ramp lane down, but it leaves the right lane virtually empty.

So with the “delayed merge” instead of slowing down and moving left I just stay in the right lane and cruise by everyone in the left lane. By the time I reach the actual merge point (extended over about half a mile), everybody has started to move left anyway and I can just move over with them.

Whenever I do this with a passenger in the car I notice they tend to get a bit uneasy. “What’s going to happen when we have to move over? Aren’t we going to get squeezed into the right-hand guard rail?” Happily after doing the delayed merge hundreds of times I have yet to take out my first guard rail.

I suppose if too many drivers were to use the delayed merge it might be a bit chaotic at the merge point, and it might be more dangerous too, since by this time cars are moving more quickly.

But as Vanderbilt demonstrates over and over in his book, you can’t take an assumption like that to the bank. You simply cannot say with any certainty how drivers will behave in an unfamiliar situation. There are lots of things the traffic engineers can predict about driver behavior. And lots of things they can’t predict.